setup-k8s-collection

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). In SKILL.md, the workflow instructs the user to paste back untrusted kubectl get pods output wrapped via wrap (e.g., kubectl get pods -n observe | wrap "kubectl-pods-observe"), which is free text ultimately ingested by the agent/LLM and could be seeded by cluster workloads.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The Cross-platform additions explicitly instruct the agent at runtime to fetch the docs pages from https://docs.observeinc.com<docs_anchor> (and use those pages' YAML as the source of truth to build/merge values), so remote content from docs.observeinc.com is fetched during runtime and directly controls the agent's generated instructions (https://docs.observeinc.com<docs_anchor>).